DE69024440T2 - METHOD FOR SPECIFICALLY REPLACING A GENE COPY, WHICH IS IN A RECEPTOR GENE, BY GENETEGRATION - Google Patents

Abstract

In a process for specific replacement of a gene, esp. by DNA targetting, the DNA being inserted consists of a part of a gene which can be made functional (or more efficient) when combined with complementary to form a complete recombinant gene in the genome of a eukaryotic cell. The new features are that (1) the insertion site is in a chosen receptor gene contg. the complementary gene; (2) eukaryotic cells are transfected with a vector contg. an insert consisting of the insert DNA and 2 flanking sequences homologous to two genomic sequences adjacent to the insertion site; (3) the inserted DNA is heterologous with respect to the receptor gene and (4) the flanking sequences make up complementary DNA so that, by homologous recombination with the corresp. sequences of the receptor gene, a complete gene is reconstructed in the genome of the eukaryotic cell. Also new are (1) plasmids for this process; (2) transformed eukaryotic cells and (3) transgenic animals in which one copy of a gene (oresent at least twice) has been inactivated by insertion of a different gene.

Description

The invention relates to a method for the specific exchange of a copy of a gene that is present in the genome of a eukaryotic recipient organism by integrating a gene that differs from the inactivated gene. Preferably, the recipient gene is present in at least two copies in the transfected host cell. The recipient gene is defined such that it is the gene at which the insertion of the other gene takes place.

In particular, the invention relates to the production of transgenic animals into which the foreign gene has been deliberately introduced, ensuring both the maintenance of the normal genetic functions of the animal and the expression of the foreign gene under the control of endogenous promoters.

By "other gene or foreign gene" is meant any nucleotide sequence that corresponds to all or part of a "foreign gene or other gene" of the recipient gene, as it is normally found in the genome (RNA or DNA), or that corresponds to an artificially modified sequence of the normal gene or to a fragment of this sequence.

The invention also relates to the process for producing the transgenic animals.

In the production of transgenic animals, the conventional methods used to introduce heterologous DNA sequences into germline cells do not allow control over the site of integration of the foreign gene in the genome, nor control over the number of copies so introduced. Integration of the foreign gene occurs randomly, and generally several copies of the gene integrate simultaneously, sometimes in a tandem head-to-tail arrangement, with the site of integration and the number of copies integrated varying from one transgenic animal to another.

It can happen that endogenous cellular genes present at the insertion site are inactivated without this being easily detectable due to the numerous random insertions. If the product of these genes is important for the development of the animal, this will be seriously disturbed. In addition, a random insertion of the foreign gene can occur at a site that is not suitable for the expression of the gene. Furthermore, the fact that the insertion site and the number of Insertions vary from animal to animal, making the interpretation of expression studies extremely difficult.

A major problem encountered in the creation of transgenic animals is achieving expression of the foreign gene. Generally, two types of experiments have been carried out in mice.

The genes introduced into the germ line are.

- either "complete" genes, comprising coding sequences flanked by their own regulatory sequences;

- or composite genes, formed from the coding sequence of one gene fused to the promoter sequence of another gene, the two fragments sometimes originating from two different animal species.

It was thus possible to confirm that the specificity of gene expression in this or that tissue is determined by their regulatory sequence(s).

The choice of the appropriate promoter for expression of the foreign gene in the transgenic animal is therefore of paramount importance.

On the other hand, directed mutagenesis of murine genes in embryonic stem cells has recently been realized using a technique of "gene targeting" (Thomas et al., 1987; Thompson et al., 1989).

In the first case, the murine HPRT gene was mutated by insertion and replacement, and in the second case, a mutated HPRT gene was corrected. Thomson et al. performed their experiments until chimeric mice were obtained and found the transmission of the genetic modification in the germline cells.

According to each of the references cited, the precise integration site was specifically determined by homologous recombination between, on the one hand, exogenous sequences containing the mutation or correction enclosed in a vector under the control of an exogenous promoter and, on the other hand, their genomic homologous equivalent. In view of this fact, it should be noted that the authors of the prior art carried out their experiments on a specific gene (HPRT) whose activation by mutation is accompanied by a recognizable phenotype. (i) The directed mutation described by Thomas et al. caused the inactivation of the HPRT gene and consequently the disappearance of the recognizable phenotype normally associated with HPRT. The selection gene NeoR under the control of a TK promoter was then introduced into the DNA insert to allow the selection of the transformants. It should be noted that the experiments described in the prior art involved selection either by the recipient gene (eg HPRT) or by the insertion gene (eg NeoR). The insertion site and/or the type of inserted gene is thus limited to genes that confer a selectable trait.

Furthermore, in the state of the art, the exogenous sequences on the vector now serve simultaneously to determine the integration site and to introduce the modification. After homologous recombination, the modified gene is always in its normal genetic environment.

It is recalled that a problem encountered during the production of transgenic animals is the risk of inactivating an endogenous cellular gene located at the insertion point of the foreign gene.

Depending on the function of the product of the inactivated gene, such inactivation may result in significant physiological or morphological disturbances in the transgenic animal or may even prevent its survival.

In contrast, inactivation of a gene could be considered beneficial if the gene in question encodes a receptor for a virus or other infectious agent.

The inventors have investigated the possibility of avoiding the disadvantages described above and, in some cases, associated with the possible inactivation of one or more endogenous cellular genes with an important function during the production of transgenic animals.

The invention relates to a method for specific exchange, in particular by targeted introduction of a DNA, wherein the DNA insertion consists of a part of a gene which can be converted into a functional form or whose functionality can be increased when it is recombined with a complementary DNA in order to form a complete recombinant gene in the genome of a eukaryotic cell, which is characterized in that

- the site of insertion is in a selected gene, the recipient gene, and contains the complementary DNA, and in that

- eucarvonic cells are transfected with a vector that contains an insertion which in turn comprises the DNA insertion and two so-called "flanking" sequences on either side of the DNA insertion, each of which is homologous to two genomic sequences that are located next to the desired insertion site in the recipient gene,

- the DNA insertion is heterologous with respect to the recipient gene, and

- the flanking sequences are selected from those which represent the complementary DNA mentioned above and which, by homologous recombination with corresponding sequences of the recipient gene, enable the reconstitution of a complete recombinant gene in the genome of the eukaryotic cell, under the Condition that when the DNA insertion is formed from a coding sequence, it is not the coding sequence of a gene encoding a selection marker.

The invention also relates to a method for producing transgenic animals, characterized in that E.S. cells are transfected under the conditions described above and selected for the occurrence of homologous recombination in order to ensure correct integration of the foreign gene, the transfected cells are injected into embryos which are at a stage in which they are able to integrate the transfected cells (e.g. at the blastocyst stage), these are then reimplanted into a surrogate mother and the chimeric individuals obtained at the end of the pregnancy are crossed. When the E.S. cells have colonized the germ line of the chimeric animal, transgenic animals which are heterozygous for the replaced gene are obtained by crossing (F1) in the offspring.

It is also possible to introduce the gene carried by the vector according to the invention into the egg shortly after fertilization (i.e. less than 24 hours). In this way, the insertion takes place while the egg is in the unicellular stage.

The invention also relates to a plasmid suitable for effecting the targeted insertion of a recombinant gene, i.e. the insertion gene, into the genome of a eukaryotic cell, which is characterized in that it contains an insertion which itself comprises the insertion gene and two so-called "flanking" sequences on either side of the insertion gene, each of which is homologous to two genomic sequences bordering the desired insertion site in the recipient gene.

The method according to the invention now allows, thanks to the phenomenon of homologous recombination, a targeted insertion of foreign genes, in particular of coding sequences without the promoter normally associated with them, into the genome of a eukaryotic organism at a site which allows their expression under the control of the endogenous promoter of the gene at which the insertion takes place, and consequently the inactivation of the endogenous target gene.

According to a preferred embodiment of the invention, the recipient target gene is a gene that is present in the genome in at least two copies. The use of the electroporation technique (reference 11) ensures the introduction of only one copy of the foreign gene.

According to this embodiment of the invention, the targeted insertion of the gene of interest (ie the so-called insertion gene) causes the inactivation of the only copy of the endogenous cellular gene at which the insertion occurs, leaving the other copy(s) of that gene intact and functional.

In this way, the overall genetic function of the transgenic animal is not or only slightly disturbed by the introduction of the foreign gene, even if the insertion inactivates a single copy of a recipient gene that is necessary for the animal's development. Either its development is not affected by the insertion of the foreign gene, or the minor disturbances that may occur in the event of inactivation of an important gene are probably not fatal to the animal. The effects of the insertion of the foreign gene in the homozygous state could be of any nature and are observed in the second generation (F2) after crossing heterozygous individuals (F1) with each other.

In contrast, if inactivation of all copies of a gene is desired, for example in the case where the gene encodes a receptor for an infectious agent, multiple copies of the foreign gene are introduced. Control of the amount introduced can be ensured by using known techniques.

The targeted insertion of the foreign gene allows its introduction at a site where its expression is under the control of regulatory sequences of the endogenous gene at which the insertion occurs.

The method according to the invention thus allows the insertion of the foreign gene behind an endogenous promoter which has the desired functions (for example expression specificity in one or another tissue), and this optionally without inactivating the other copies of the recipient gene,

According to a particularly preferred embodiment of the invention, the DNA insertion between the flanking sequences comprises, on the one hand, a DNA sequence which is to be recombined with the complementary DNA in the recipient gene to yield a recombinant gene and, on the other hand, a sequence which encodes a selection marker which allows the selection of the transformants and a promoter which controls the expression of the selection marker, the recipient gene and the recombinant gene encoding expression products which do not confer a selectable phenotype.

In this way, the selection of transformants is completely independent of the nature of the recipient gene and the inserted gene, in contrast to the methods described up to this day, in which the inserted gene or the recipient gene had to necessarily encode an expression product that allowed the selection of transformants. The system developed by the inventors allows complete flexibility as regards the nature of the recipient gene and the inserted gene. or of the gene formed by homologous recombination. The inventors have surprisingly found that the insertion of long sequences (eg about 7.5 kb) does not affect the frequency of homologous recombination.

The effect that the insertion of the DNA sequence according to this aspect of the invention can have includes, depending on the type of inserted sequence, for example, the replacement of a coding sequence, the replacement of a regulatory sequence, the inactivation or reactivation of a gene by mutation or the improvement of the expression level of a gene. It is possible according to the invention to replace a coding section or part of a coding section with a heterologous sequence that begins at the initiation codon of the exchanged gene so that the expression of the inserted gene completely replaces the expression of the exchanged gene. This avoids the formation of fusion proteins, which could be undesirable in a transgenic animal.

According to this embodiment of the invention, the DNA insert between the flanking sequences may carry a heterologous coding sequence without the promoter, wherein the coding sequence is not a gene encoding a selectable marker. The DNA insert may further carry, downstream of the coding sequence and still between the flanking sequences, a gene encoding a selectable marker and associated with a promoter allowing its expression in the target cell.

In this way, the heterologous coding sequence can be inserted behind an endogenous promoter that has the desired properties, such as a certain expression specificity or a certain transcription pattern, etc., whereby the selectability of the transformed cells is completely independent of the expression of the heterologous coding sequence. This type of construction allows, for example, the selection of the transformants even if the gene replaced by the heterologous coding sequence is not normally expressed in the target cells. This is particularly important when producing transgenic animals from an E.S. cell (embryonic stem cell), since a significant proportion of the genes remain inactive until a more advanced stage of the animal's development. The Hox 3.1 gene is an example of such a gene. On the other hand, if the coding sequence encodes an easily detectable protein, for example β-Gal, the evolution of the transcription pattern of the endogenous exchanged gene can be followed. The vector pGN is an example of this type of construction.

According to another embodiment of the invention, the inserted DNA may carry a foreign regulatory sequence. The insertion site and consequently the flanking Sequences are selected according to the desired goal, ie, either the insertion of the foreign regulatory sequence to give a "double promoter effect" with the endogenous regulatory sequence, or the replacement of an endogenous promoter with the foreign promoter. The coding sequence under the control of the regulatory sequence may be endogenous.

Another possibility is the targeted insertion of a foreign DNA that simultaneously contains a regulatory sequence and a coding sequence. It is possible that the regulatory sequence is the one that is naturally associated with the coding sequence.

The method according to the invention uses a vector which contains two "flanking" sequences on either side of the foreign gene. These flanking sequences are at least 150 base pairs long and are preferably shorter than the recipient gene. These two flanking sequences must be homologous to two genomic sequences adjacent to the desired insertion site. The flanking sequence of the vector which is present upstream of the foreign gene to be introduced is normally homologous to the genomic sequence which is located on the 5' side of the insertion site. Likewise, the flanking sequence of the vector which is located downstream of the foreign gene is normally homologous to the genomic sequence which is located on the 3' side of the insertion site.

It is possible to introduce "intercalating" (intervening) sequences between one or other of the flanking sequences and the foreign gene, for example sequences that allow the selection of transformants, markers, sequences that allow the cloning of the vector, etc.

However, the position of these intercalating sequences with respect to the foreign gene must be chosen so that it does not hinder the expression of the foreign gene, in particular the foreign coding DNA sequence under the control of the endogenous promoter or, conversely, the endogenous coding DNA sequence under the control of foreign regulatory elements introduced by the insertion sequence.

Despite the presence of flanking sequences that support homologous recombination, it is possible that a certain number of integrations occur randomly. To confirm that the targeted insertion has occurred at the target site and not at another location, the polymerase chain reaction (PCR) technique (see reference 10) is used to amplify the DNA sequences of the site where the insertion should have occurred. only the clones transformed as a result of homologous recombination are selected.

The flanking sequences of the vector are of course selected according to the desired insertion site so that homologous recombination can take place. Optionally, the flanking sequences may contain sequences corresponding to the endogenous promoter and/or sequence modifications preceding the initiation codon in order to improve the level of translation (upstream sequences) and sequences corresponding to the termination sequences, in particular polyadenylation sites (downstream sequences).

The insertion gene can be any gene of interest. Non-limiting examples include the lacZ gene (as in the model described later), the genes encoding interleukin or interferon, the genes for a receptor, for example the retinoic acid or β-3 adrenergic or HIV receptor, and genes known to be associated with certain diseases, e.g. myopathy, etc.,...

According to a preferred embodiment of the invention, the eukaryotic cells are embryonic stem cells (see references 14 and 15).

Indeed, a mutated E.S. cell can be injected into an early embryo, which, after reimplantation, can be born in a chimeric form. When the germ line is colonized by the mutated cell, the chimeric animal transmits the mutation to its offspring. As a result, the effects of this mutation in the homozygous state can be observed in certain individuals on their development, behavior, metabolism, pathology, etc., ...

Figure 1 shows the plasmid pGN.

Figures 2a and b show the molecules pGMA and pGMD, respectively, constructed from the plasmid pGN for the Hox-3.1 gene. These plasmids are mutagenesis plasmids. Shown are two parts of the coding section of the Hox-3.1 gene, on chromosome 15, with the "homeo" box in black. The corresponding Hox-3.1 sequences were cloned into the plasmid pGN.

(A: polyadenylation signal Enh/Pro: enhancer-promoter)

07 and 08 indicate the two oligonucleotides used in the PCR.

Figures 3 to 6 show the plasmids used in the construction of pGN.

Figure 7 shows the detection of homologous recombination using the polymerase chain reaction (PCR) technique on transfected E.S. cells.

Figure 8(a) and (b) shows the Southern analysis of single positive clones (L5 and F2) and E.S. cells (C.C E.).

The method according to the invention has significant industrial applicability and can vary depending on the nature of the foreign gene introduced.

Mammalian genetics will progress considerably thanks to the recent possibility of specific mutagenesis of any gene, thus making it possible to better define its role. This technique, which uses homologous recombination and E.S. cells, will provide valuable information on oncogenes, growth factors, transcription factors, etc., as well as on genes that are currently the subject of basic or applied research. An important area of application for medical research is the possibility of reproducing a human disease whose genetic origin is known (certain human pathological disorders such as Duchesne's myopathy), in order to better study the mechanisms and find a therapy.

When applying the method according to the invention, a gene known to be responsible for a particular disease is specifically inserted into the genome of an E.S. cell. The transgenic animal subsequently produced represents a useful model for this disease.

Where appropriate, and as described above, normal genetic functions may be significantly maintained despite the insertion of the foreign gene.

Another application of the method according to the invention is to introduce an insertion gene which is easily detectable, e.g. the lacZ gene, and which can thus play a role as a cellular marker. In this way, lineage studies, e.g. in competing animals, can be facilitated and the breed can be traced.

The insertion of the lacZ gene as an insertion gene also allows studies on the promoter. Thanks to the possibility of detecting β-galactosidase activity, the activity and specificity of different endogenous promoters can be studied by inserting them into different sites in the same cell type or in different cell types. The same studies could also be carried out on a whole organism during development or in the adult stage using the techniques of chimeric or transgenic animals.

The inventors have surprisingly found that the frequency of homologous recombination is not affected by the insertion of fragments of considerable size, for example lacZ. This observation suggested to the inventors that the technique of homologous recombination is well suited for the insertion of other heterologous genes which are of considerable size.

Thanks to the possibility of modifying the genome of an animal, the method according to the invention can also be used as "gene therapy". The most obvious Uses consist in inactivating genes encoding receptors for infectious (viruses or bacteria) or toxic agents. If this mutagenesis were to prove lethal, the lost function would have to be restored without restoring sensitivity to the harmful agents. A modified gene encoding such a receptor could be reintroduced into the mutated cell, unless the modification could be brought about by homologous recombination. This modification of the genetic heritage would confer on the animal immunity to the disease in question.

This protocol can also be used for autotransplants. Sick or healthy cells taken from a patient could be treated and immunized and then reimplanted into the same individual.

The technique of the invention is also suitable for activity studies of pharmaceutical products suspected of having activity on the expression products of a pathological gene associated with a disease. In this case, the insertion gene consists of the pathological gene and the pharmaceutical product is administered to the transgenic animal to evaluate its effect on the disease.

The invention is explained with reference to the plasmid pGN and its use in the targeted insertion of a foreign gene (lacZ encoding the enzyme β-galactosidase of E. coli) into the genome of a mouse E.S. cell. The lacZ gene was chosen because its expression can be easily detected and serves for ease of explanation.

The coding region of the E. coli β-galactosidase enzyme (lacZ; 1-3057), fused to a genomic sequence (7292-3) of the mouse gene Hox.3-1 (Ref. 1), begins with the start codon of this gene. Indeed, the sequence preceding the start codon of Hox-3.1 is identical to the consensus sequence observed in vertebrates (Ref. 2), thus allowing a better translation level of β-galactosidase in vertebrate cells. The lacZ gene, as with most eukaryotic genes, is followed by a polyadenylation signal, for example from the SV40 virus, to stabilize the messenger RNAs.

The activity of E. coli β-galactosidase, which is functional in eukaryotic cells, can be detected in several ways. After fixation, cells expressing the lacZ gene acquire a blue color in the presence of X-Gal, which is a substrate of β-galactosidase (reference 3). A new substrate DGF (di-β-galactopyranosidofluorescein) allows the detection and determination of β-Gal activity, while keeping the cells alive (reference 4). The cells that lacZ express a fluorescent compound and can be isolated using a cell sorting device or FACS ("fluorescence-activated cell sorter").

The transcription unit of the neomycin resistance gene is derived in large part from the plasmid pRSV neo (reference 5). The LTR (long terminal repeat) of the Rous sarcoma virus represents very strong activator and promoter sequences in numerous eukaryotic cells (reference 6). The bacterial transposon Tn5 provides a promoter active in E. coli and the coding section of the enzyme phosphotransferase (reference 7), followed by the polyadenylation signal of the SV40 virus. The same gene under the dual control of the RSV and Tn5 promoters can confer neomycin or kanamycin resistance in bacteria and G418 resistance in eukaryotic cells.

Through the action of a simple point mutation, unit B of the activator sequences (enhancers) of the PyEC F9.1 strain of the polyoma virus became much more active in various cell types, in particular in embryonic carcinoma (EC) cells (reference 8). Two copies of this enhancer PyF9.1 were inserted in tandem into the plasmid pGN upstream of the LTR-RSV and in the orientation of the "late promoter" of the regulatory region of the polyoma virus.

To improve the translation efficiency of the phosphotransferase, the sequence preceding the initiation codon was modified by oligonucleotide mutagenesis. The sequence T T C G C A U G became G C A C C A U G, which corresponds much better to the consensus sequence for translation initiation in vertebrates (reference 2).

The improvements introduced to the transcription unit of the neomycin resistance gene could be assessed by transfection of mouse embryonic stem cells (ESCs). At the same molarity relative to the plasmid, a construction with the PyF9.1 enhancers yielded 7.5-fold more G418-resistant clones than pRSV neo and 2-3-fold more than pMC1 neo, as described by Capecchi et al. (ref. 13). The number of clones was again increased 60-fold, i.e. 450-fold relative to pRSV neo, by modifying the translation initiation sequence. Homologous recombination can be a fairly rare event under the experimental conditions used (e.g. 1/1000 for HPRT, ref. 13). A vector that has an increased selection efficiency is therefore very useful, especially since the electroporation conditions predominantly lead to the integration of a single copy.

The plasmid pGN also contains a bacterial replication origin of the type colE1, pBR322, which allows cloning and preparation in E. coli.

Finally, an in vitro synthesized multiple cloning site (MCS) containing cleavage sites unique to pGN was inserted upstream of lacZ to facilitate the uses of this plasmid.

The "flanking" plasmid sequences that induce homologous recombination were added to the ends of the plasmid pGN after linearization of the plasmid upstream of lacZ through an MCS cleavage site (see Figure 2). In the present case, the flanking sequences chosen are homologous to chromosomal sequences derived from Hox-3.1 that will later be used in homologous recombination.

Figure 2 shows the molecule constructed from the plasmid pGN in relation to the Hox-3.1 gene. In this case, recombination between the plasmid sequences and chromosomal sequences of Hox-3.1 leads to an insertion at the beginning of the coding region of this gene, thus leading to its complete inactivation.

The pGN plasmid has several advantages for this procedure, which can be applied to any gene. Since the homologous recombination event can be quite rare (of the order of 1 in 1000 non-homologous integrations), it is necessary to be able to analyze a large number of clones whose G418 resistance is strong enough to establish itself in the case of insertions in any part of the genome. The modifications made to the transcription unit of the phosphotransferase meet these problems perfectly. The homologous recombination mutagenesis procedure corresponds to the inactivation of a gene by an insertion or a substitution, but the pGN plasmid has the additional advantage of being able to replace the expression of β-galactosidase with that of the mutated gene. Finally, MCS facilitates the cloning of genomic fragments.

Examples: I - Construction of the plasmid pGN

The intermediate plasmids are numbered according to their level.

1st stage:

Insertion of a XhoI cleavage site into the BglI cleavage site of pRSV neo

Insertion of a XhoI linker into the BglI cleavage site of pRSV neo, filled in by the Klenow fragment of the DNA polymerase from E. coli.

2nd stage:

Insertion of a ClaI cleavage site into the NdeI cleavage site of plasmid p1

Insertion of a ClaI linker into the NdeI cleavage site of p1, filled in by the Klenow polymerase.

3rd stage:

Insertion of the enhancer PyF9.1 into the ClaI site of plasmid p2

Insertion of the enhancer PyF9.1 PvuII-PvuII, isolated at a single AccI cleavage site, into the ClaI cleavage site of p2. Selection of a clone containing two enhancers oriented towards the late promoter.

4th stage:

SmaI-Hpa I deletion of plasmid p3

The two enzymes that result in blunt ends can be directly ligated. This deletion removes the intron of the SV40 t-antigen, which is not very useful, and significantly reduces the size of the transcription unit of the phosphotransferase.

5th level:

Insertion of a XhoI cleavage site into the BamHI cleavage site of pCH110

Insertion of a XhoI linker into the BamHI cleavage site of the plasmid pCH 110 (Pharmacia), filled in by Klenow polymerase.

6th level:

Insertion of 3'-lacZ-polyA SV40 into plasmid p4

The 3' part of the coding region of β-galactosidase, followed by the polyadenylation signal of the SV40 virus, is isolated from plasmid p5 through the XhoI-AatII cleavage sites and cloned into plasmid p4 at the same cleavage sites.

7th level:

Insertion of 5'-lacZ into the vector KS-

The 5' part of the coding region of β-galactosidase is isolated from the plasmid pMC 1871 (Pharmacia) through the PstI-SacI cleavage sites and cloned into the vector KS- (Stratagene) at the same cleavage sites.

8th level:

Fusion of a genomic Hox-3.1 sequence with 5'-lacZ

A genomic sequence of the Hox-3.1 gene cloned into the vector KS- is purified by successive digestions by the enzyme SacI, then by mung bean nuclease and finally by the enzyme ApaI. This insertion is fused to the 5' end of the coding portion of β-galactosidase by cloning into the plasmid p7 digested by ApaI-SmaI. The protein thus fused contains the translation start codon of the Hox-3.1 gene followed by the coding portion of β-galactosidase (subsequent confirmation by sequencing). Mung bean nuclease

9th level:

Insertion of Hox-3.1-5'-lacZ into plasmid p6

The fusion Hox-3.1-5'-lacZ is isolated from the plasmid p8 through the cleavage sites ApaI-SacI and cloned into the plasmid p6 at the same cleavage sites. This cloning results in the reconstitution of the coding section of the β-galactosidase in its full length.

10th level:

Insertion of the NeoR gene into the vector KS+

The neomycin resistance gene (bacterial promoter and coding section of the phosphotransferase) is isolated from pRSV neo through the cleavage sites HindIII-EcoRI and cloned into the KS+ vector (Stratagene).

11th level:

Mutagenesis of the initiation sequence of NeoR in p10

The translation initiation sequence of the phosphotransferase is modified to be identical to the consensus sequence observed in vertebrates, allowing a higher level of translation initiation and thus increased G418 resistance in mammalian cells. The modification also created an ApaLI cleavage site that allows control of the success of the mutagenesis.

An oligonucleotide (CTTGTTCAATCATGGTGCACGATCCTCA) containing a mismatch region with the pRSV neo sequence (underlined) is synthesized (Gene Assembler, Pharmacia), then phosphorylated by the bacteriophage T4 polynucleotide kinase. A single-stranded template of plasmid p10 is prepared from the f1 origin of the KS+ plasmid and hybridized with the mutagenesis oligonucleotide. The second strand is synthesized and repaired by Klenow polymerase md the bacteriophage T4 DNA ligase. After transformation of bacteria, the mutated clones are screened using the 32P-labeled oligonucleotide. Mutagenesis was confirmed by digestion with Apa LI as well as by sequencing.

12th level:

Exchange of the initiation sequence in the plasmid p9

A fragment containing the modified translation initiation sequence of the neomycin resistance gene is isolated from plasmid p11 by the enzymes HindIII-EagI and cloned into plasmid p9 at the same cleavage sites.

13th level:

Insertion of the multiple cloning site into the plasmid p12

Two complementary oligonucleotides are synthesized (Gene Assembler, Pharmacia), then phosphorylated. After pairing, the MCS is cloned via its cohesive ends into the ApaI-SacII cleavage sites of the plasmid p12

The multiple cloning site was also confirmed by sequencing.

II - Addition of the "flanking" sequences to the ends of the plasmid pGN linearized at an MCS cleavage site upstream of lacZ'

The flanking sequences used were chosen according to the desired insertion site (e.g. Hox-3.1, see Figure 2a and b, pGMA and pGMD).

In constructing the mutagenesis plasmid pGMD, two DNA arms homologous to the Hox-3.1 locus were cloned into the ApaI-NsiI and NsiI-SacII cleavage sites of the vector pGN. The 5' arm begins at the SacII cleavage site (CCGCGG) at nucleotide 219 of the Hox-3.1 cDNA c21. This fragment extends 6.8 kb in the 5' direction to the first BamHI cleavage site. The 3' arm begins at the ApaI cleavage site (GGGCCC) at Nucleotide 885 of cDNA c21. This fragment extends 1.5 kb in the 3' direction to the first PstI cleavage site. An NsiI linker was inserted into the BamHI cleavage site of the 5' fragment and into the PstI cleavage site of the 3' fragment. The 5' and 3' arms were cloned into the pGN vector into the NsiI-SacII and ApaI-NsiI cleavage sites, respectively. The sequence of cDNA c21 of Hox-3.1 has been published (Reference 1).

The mutagenesis plasmid was linearized by digestion with NsiI before electroporation of E.S. cells. Its ends form the two genomic arms that were cloned into the ApaI-NsiI and NsiI-SacII cleavage sites of the vector pGN.

The plasmid pGMD does not have a polyadenylation signal after the resistance gene, but does have an AU-rich sequence that is responsible for selective degradation of the mRNA that is inserted into the intron sequence of Hox-3.1 of the plasmid.

Another mutagenesis plasmid, pGMA, has the same structure as pGMD, but contains the SV40 transcription polyadenylation and termination signals and lacks an AU sequence downstream of the Neor gene for mRNA degradation. These modifications are expected to reduce the amount of Neor transcripts in clones resulting from random integration. Conversely, clones resulting from homologous recombination events between pGM and a Hox-3.1 locus should have unaltered growth during selection for G418, with the AT sequence for mRNA degradation eliminated by the recombination event itself or spliced out with the Hox-3.1 intron.

In the following experimental steps, the protocol described by Thompson et al 1989 was used to generate the chimeric animals.

III - Transfection of embryonic mouse cells

The method described by Thompson et al. in 1989 was used to transfect embryonic mouse cells. The use of electroporation technology ensures the introduction of a single copy of the foreign gene (lacZ) per cell. After transfection, several clones expressing β-galactosidase were isolated

The mutagenesis plasmids pGMD and pGMA were linearized and introduced into the E.S. cells by electroporation to favor the insertion of only one copy into the genome (reference 11).

The sufficient transfections were performed to compare the efficiency of the plasmids pGMA and pGMD to insert specifically into Hox-3.1 (see Table I) Table 1 Homologous recombination in the Hox 3.1 gene Experiment Mutagenesis plasmid No. of the analyzed group Number of clones forming the group Number of positive PCR results

The E.S. cell line "C.C.E." (Reference 16) was maintained by continuous culturing on fibroblast feeder layers (Reference 17). For experiments I and II, 1.5 x 10⁷ E.S. cells were electroporated in 1.5 ml HeBS at 200 V with 40 mg linearized plasmid (Reference 11), and then plated on four culture dishes (diameter 100 mm). For experiment III, the shock was performed under the same conditions, but a quarter of the cells were plated on four 24-well plates. The following day, 250 µg ml-¹ G418 was added. Each transfection resulted in approximately 2400 clones with pGMA and 1000 clones with pGMD.

The mean number of clones of G418-resistant E.S. cells in each group is given in Table 1, as is the number of those giving a positive result by the PCR technique. A positive result means that a 1.6 kb band can be observed on an agarose gel stained with ethidium bromide (see Figure 7). The number of clones giving a positive signal after Southern analysis of the PCR mixture and hybridization with a specific probe that does not contain the starter sequences is given in parentheses (Figure 8).

Detection of homologous recombination using PCR

The PCR was carried out on 105 cells from a total of 250 clones of transfection II (see lane D of Figure 7). In the other lanes, four clones of transfection III were examined together as a mixture of about 4 x 5000 cells. The starter sequences 07 and 08 used in the PCR comprise the sequence 3'-Hox-3.1 of the mutagenesis plasmid (Figure 2). The 1.6 kb fragment covering this 3' sequence can only be amplified in the case of homologous recombination. Lanes 2, 3 and D show positive results.

The DNA of the ES clones was prepared on a filter replica using the boiling proteinase K digestion-boiling method (reference 18). 40 Amplification cycles (40 s at 94°C, 1 min at 60°C, 7 min at 72°C) were performed in a 100 µl reaction mixture containing 67 mM Tris-HCl (pH 8.6), 16.7 mM (NH₄)₂SO₄, 6.7 mM MgCl₂, 10 mM 2-mercaptoethanol, 0.01% (w/v) gelatin, 200 µM dATP, dTTP and dCTP, 100 µM dGTP, 100 µM 7-deaza-dGT, 600 ng of each starter (07. AACTTCCCTCTCTGCTATTC and 08. CAGCAGAAACATACAAGCTG) and 3 U Taq polymerase (Perkin Elmer Cetus) covered with 100 µl paraffin. Half of the reaction mixture was applied to a 0.7% agarose gel stained with ethidium bromide. The length marker is an EcoRI + HindIII digest of the λ-DNA.

Southern-Analvse

Three independent clones of E.S. cells containing the Hox-3.1 mutation (identified by PCR) were isolated from the group of positive clones using pipettes. Their DNA was examined by Southern analysis after digestion with the restriction enzymes indicated in Figure 8 to confirm the specific insertion and to distinguish between the recombinant and wild-type loci. Two different probes were used in the analysis of the 3' end of the Hox-3.1 loci in the mutated clones and in the non-mutated E.S. cells, which served as a control (Figure 8c). The first probe "a" was contained in the Hox-3.1 sequences of the mutagenesis plasmid and showed the number of integrations of the vector and their physical linkages. One of the three recombinant clones further contained a randomly integrated copy of the plasmid (Figure 8a, clone F2). The second probe "b", which was not included in the mutagenesis vector, discriminated between the recombinant and wild-type Hox-3.1 alleles (Figure 8b). The recombinant Hox-3.1 locus showed the hybridization pattern expected from restriction maps of the mutagenesis vector and the intact site with the two probes. Furthermore, the presence of two recombination domains in the 3' arm of the vector was confirmed by the presence or absence of the AT sequence in the recombinant Hox-3.1 locus (e.g. Figure 8, clone L5). The 5' end of the Hox-3.1 locus was also examined for a homologous recombination event. Restriction enzymes that do not contain cleavage sites in the 6.8 kb 5' Hox-3.1 sequence of the mutagenesis vector were used to digest the DNAs of the recombinant clones. These DNAs were then subjected to pulse-field gel electrophoresis to distinguish the fragments with increased molecular weight. Southern analysis of this gel also showed the correctly recombined alleles and the wild-type alleles of Hox-3.1 using a probe that contained a sequence upstream of the mutagenesis plasmid.

Southern analyses showed that one allele of the Hox 3.1 gene was recombined as intended. The homologous recombination was equivalent to a double crossing-over between the genomic arms of the mutagenesis plasmid and the homologous chromosomal sequences (Figure 2).

In the recombinant clones, the lacZ gene was under the control of Hox-3.1 promoter and regulatory sequences upstream of the AUG codon, but the 3' mRNA maturation signals were from SV40. In these recombinant clones, lacZ expression was undetectable by X-Gal staining, consistent with the absence of Hox-3.1 transcription in the E.S. cells. This was demonstrated by RNase protection analysis. β-Gal activity could be induced in certain cells after 3 or 4 days of culture in the presence of 5x10-7 M retinoic acid, conditions known to induce Hox-3.1 transcription (ref. 19).

Using the mutagenesis vector pGMA, which shows complete homology to the Hox-3.1 locus over 8.3 kb, a 120 bp fragment was replaced by a 7.2 kb insertion. The frequency of this targeted exchange (1/900) is comparable to that recently obtained with HPRT (1/1000) (ref. 13) or with En-2 (1/260) (ref. 20), although in the latter cases the inserted heterologous fragment was much smaller in size (1.1 and 1.5 kb, respectively). Surprisingly, it was found that a greatly increased homologous recombination frequency (1/40) can be achieved with the vector pGMD. The removal of the 3' maturation signals of the mRNA and the addition of the degradation sequence of the mRNA to the neomycin resistance gene resulted in a reduction in the total number of G418-resistant clones by a factor of 2.4 (Table I). The ratio of specific insertion was increased by about 10-fold (900/40). The mechanism of homologous recombination itself must have been influenced in the experiments with pGMD. A possible explanation of these results is that an AT sequence of 51 pb could form an open loop in vivo in the mutagenesis plasmid due to its reduced melting temperature. If the Hox-3.1 sequences adjacent to pGMD on either side of the AT region can be influenced by this opening, they could react more efficiently with the chromosomal Hox-3.1 locus in the single-stranded state. The model of mitotic recombination in yeast suggests that it is initiated by such strand exchange, although the mechanism of homologous recombination in the more complex eukaryotes remains unknown.

Figure 8 shows the results of Southern analysis performed on individual positive clones (L5 and F2) and on E.S. cells (C.C.E.).

The probes used hybridize only with the Hox-3.1 sequences contained in the vector (a) or eliminated from the mutagenesis vector (b). The hybridization pattern of the recombinant Hox-3.1 locus (open triangles) is clearly different from that of the wild-type locus (black triangles). The stars indicate the hybridization bands of a copy of the plasmid that has integrated randomly. The size marker is an EcoRI + HindIII cleavage of the λ DNA.

Figure 8(c) shows restriction maps of recombinant (rec.) and wild-type (wt.) Hox-3.1 alleles. The parts of the mutagenesis vector and the Hox-3.1 locus are indicated with the same symbols as used in Figure 2. In this case, the AT sequence was integrated by homologous recombination. The vertical arrow indicates the 3' end of the mutagenesis plasmid. The location of probes "a" and "b" used in Southern analysis is also indicated. In Figure 8, the following abbreviations were used: B, BamHI; D, DraI; E, EcoRI; H, HindIII; S, SalI; X, XhoI.

IV - Production of chimeric embryos

Microinjection into blastocysts was performed with two recombinant E.S. clones containing one intact Hox 3.1 allele and one recombinant allele, whereby these clones did not contain another copy of the mutagenesis plasmid. The karyotypes of these cells were normal.

10 to 15 mutant cells were microinjected per blastocyst. After reimplantation in surrogate mothers, embryos were collected at 9.5, 10.5 and 12.5 days post-conception and examined for lacZ expression. The transcription pattern of Hox-3.1 at these stages had previously been determined by in situ hybridization analysis (Reference 1). Hox-3.1 transcripts are first detectable at late gastrulation and distributed to all tissues of the posterior part of the animal. Later, the distribution becomes increasingly spatially restricted and tissue-specific. At day 12.5 post-conception, transcription is localized in the cervical region of the neural tube, at the level of the heart. Thus, the distribution of Hox-3.1 transcription is modified during embryogenesis. The postconception day 10.5 stage appears to be a transitional period during which transcription occurs simultaneously in the two posterior regions and the cervical neural tube.

In the chimeric embryos of 9.5 and 10.5 days after conception, the caudal part in the posterior embryo showed intense β-Gal activity, while the marker was never detected in the anterior thorax region or in the head (Figure 9a). In the posterior region, β-Gal-stained cells were found in all tissues and in all embryonic layers. Between the two primordia giving rise to the limbs, the stained cells were distributed in restricted zones, in the upper ectoderm (Figure 9b), as well as in the posterior regions (Figure 9c) and in the form of narrow lines or rays in the neural tube (Figure 9b). These rays showed an irregular and asymmetric distribution on the neural tube envelope. Transcription of Hox-3.1 was not detected in the thin layer of cells towards the end of the neural tube. These cells were probably not resistant to the treatments applied during in situ hybridization. It was observed that the cells of the neural ectoderm very soon become involved in different parts of the nervous system and migrate in a radial direction, corresponding to narrow lateral movements (Reference 21). These results are therefore consistent with this observation.

Thus, the expression of lacZ correctly detected the first part of the transcription of the homolog Hox-3.1, i.e. in all tissues of the caudal regions of the embryos of 9.5 and 10.5 days after conception, and provided new information regarding the transcription mode of Hox-3.1.

In contrast, expression of lacZ was not observed in the cervical regions of the neural tube of 12.5 day chimeric embryos, nor in the anterior region of 10.5 day embryos; this result was not expected from the in situ hybridization studies. The later phase of Hox-3.1 transcription, which was observed in very localized zones of the neural tube from day 10.5 onwards, was not detected by β-gal activity. A possible explanation for this result would be that, although expression of lacZ occurs under the control of the Hox-3.1 promoter, the 3' sequences of Hox-3.1 are absent from the reporter gene. It is possible that sequences 3' of the AUG initiation codon of Hox-3.1 have an influence on the late expression of Hox-3.1 in the anterior [)domain. A gene dosage effect could also explain this result. In Drosophila, the self-activation of various homeogens has been demonstrated genetically or suggested by the formation of complexes between the DNA and the homeobox proteins.

If the late part of Hox-3.1 transcription in the neural tube is maintained by a similar mechanism, inactivation of one allele could have a dominant effect on the cells of the neural ectoderm. Since only one allele produces the Hox-3.1 protein, the activation signal should be distributed between the two promoters. Reducing self-(in)activation at the two loci could also lead to a complete arrest of transcription initiation. This could explain why no expression of lacZ was observed at all in the cervical region of the neural tube of 10.5 and 12.5 day embryos.

V - Transmission of the modification in the germline cell line: production of transgenic animals

The effects in F1 and F2 of the modification introduced by the targeted insertion were observed after propagation of the chimeras. Transmission of the modification in the germline cell line was established.

References

1. Le Mouellic, H.1 Condamine, H. et Brület, P. (1988). Genes & Development. 2, 125-135.

2. Cavenir, D.R. (1987). Nucleic Acids Res. 15, 1353-1361.

3. Sanes, J.R. Rubenstein, J.L.R. and Nicolas, J.F. (1986) EMBO J. 5, 3133-3142.

4. Nolan, G.P., Fiering, S., Nicolas, J.F. et Herzenberg, L.A. (1988) Proc. Natl. Acad. Sci. US 85, 2603-2607.

5. Gorman, C., Padmanabhan, R. and Howard, B.H. (1983). Science. 221, 661-553.

6. Gormann, C.M., Merlino, G.T. Willingham, M.C. Pastan, I. et Howard, B. H. (1982) Proc. Natl. Acad. Sci. U.S. 79, 6777-6781.

7. Southern, P.J. et Berg, P. (1982) J. Mol. Appl. Genet 1, 327-341.

8. Herbomel, P., Bourachot, B. et Yaniv, M. (1984). Cell. 39, 653-662.

9.Robertson, E.J. (1987). Teratocarcinomas and embruonic stem cells: A practical approach. IRL Press, Oxford.

10.Kahn, A. (1988). Medicine/Sciences 4, 515-518. 11.G. Chu, Hayakawa H., Berg. P. (1987) Nucleic Acid Research 15, No. 3, 1311-1326.

12.Thompson, S., Clarke, A.R., Pow, AM., Hooper, M.L., Melton, D.W., (1989) Cell, 56, 313-321.

13.Thomas, K.R.,Capecchi, HR., (1987) Cell, 51 503-512.

14. Evans, M.J., Kaufmann, M.H. (1981) Nature, 292, 154-155,

15.Robertson, E.J., (1986) Trends in Genetics, 9-13

16.Robertson, E., Bradley, A., Kuehn, M. & Evans, M. Nature 323, 445-448 (1986).

17.Robertson, E.J. in Teratocarcinomas and Embryonic Stem Cells (ed. Robertson, E.J.) 71-112 (IRL, Oxford, 1987).

18.Kim, H.S. & Smithies, O. Nucleic Acids Res. 16, 8887-8903 (1988).

19.Breier, G., Bucan, M., Francke, U., Colberg-Poley, A.M. & Gruss, P. EMBO J.5, 2209-2215 (1986).

20.Joyner, A.L., Skarnes, W.C. & Rossant, J. Nature 338, 153-156 (1989).

21.McKay, R. D. G. Cell 58, 815-821 (1989).

Claims (1)

1. Method for the specific exchange or specific insertion of a gene, in particular by targeted introduction of a DNA, the DNA insertion consisting of the entirety of a gene or of a part of a gene which can be converted into a functional form or whose functionality can be increased when it is recombined with a complementary DNA in order to form a complete recombinant gene in the genome of a eukaryotic cell, characterized in that - the site of insertion is in a selected gene, the selected recipient gene, and contains the complementary DNA, and in that - eukaryotic cells are transfected with a vector that contains an insertion that in turn includes the DNA insertion and two so-called "flanking" sequences on either side of the DNA insertion, each of which is homologous to two genomic sequences that are located next to the desired insertion site in the recipient gene, - the DNA insertion is heterologous with respect to the recipient gene, and - the flanking sequences are selected from those which constitute the complementary DNA mentioned above and which, by homologous recombination with corresponding sequences of the recipient gene, enable the reconstitution of a complete recombinant gene in the genome of the eukaryotic cell, provided that, when the DNA insert is formed by a coding sequence, it is not the coding sequence of a gene encoding a selection marker. 12. Method for the specific exchange or specific insertion of a gene, in particular by targeted introduction of a DNA, wherein the DNA insertion consists of the entirety of a gene or of a part of a gene that can be converted into a functional form or whose functionality can be increased when it is recombined with a complementary DNA in order to form a complete recombinant gene in the genome of a eukaryotic cell, characterized in that - the site of insertion is in a selected gene, the selected recipient gene, and contains the complementary DNA, and in that - eukaryotic cells are transfected with a vector that contains an insertion that in turn encompasses the DNA insertion and two so-called "flanking" sequences on either side of the DNA insertion, each of which is homologous to two genomic sequences that are located next to the desired insertion site in the recipient gene, - the DNA insertion is heterologous with respect to the recipient gene, and - the flanking sequences are selected from those which represent the above-mentioned complementary DNA and which, via homologous recombination with corresponding sequences of the recipient gene, enable the reconstitution of a complete recombinant gene in the genome of the eukaryotic cell, - the recipient gene is a gene that is not normally transcribed in the eukaryotic cell. 3. Method according to claim 1 or 2, characterized in that the DNA insertion comprises either a coding sequence or a regulatory sequence, - the flanking sequences are selected in such a way that, due to homologous recombination, they either inhibit the expression of the coding sequence of the entire DNA insertion under the control of regulatory sequences of the recipient gene or the expression of a coding sequence of the recipient gene under the control of the regulatory sequences of the DNA insertion. 4. Method according to one of claims 1, 2 or 3, characterized in that the DNA insertion comprises a coding sequence which has no regulatory elements, in particular no promoter of its own. 5. Method according to claim 3, characterized in that the DNA insertion comprises a regulatory sequence. 6. Method according to claim 5, characterized in that the endogenous promoter of the recipient gene is inactive in the transfected eukaryotic cells. 7. Method for the specific exchange of a regulatory sequence of a gene, the recipient gene, by targeted introduction of a DNA, the DNA insertion, whereby the DNA insertion contains a regulatory sequence, characterized in that - eukaryotic cells are transfected with a vector that contains an insertion that in turn includes the DNA insertion and two so-called "flanking" sequences on either side of the DNA insertion, each of which is homologous to two genomic sequences that are located next to the desired insertion site in the recipient gene, - the DNA insertion is heterologous with respect to the recipient gene, and - the flanking sequences are chosen in such a way that they allow the expression of a coding sequence of the recipient gene under the control of the regulatory sequences of the DNA insertion due to homologous recombination. 18. Method according to claim 7, characterized in that the endogenous promoter of the recipient gene is normally inactive in the transfected eukaryotic cells. 9. Method according to one of the preceding claims, characterized in that the recipient gene occurs in the genome of the eukaryotic cell in at least two copies. 10. Method according to one of the preceding claims, characterized in that each of the flanking sequences is longer than 150 base pairs and shorter than the recipient gene. 11. Method according to one of claims 1 to 10, characterized in that the eukaryotic cells are embryonic stem cells (ES). 12. Method according to one of claims 1 to 11, characterized in that the DNA insert is a heterologous gene or a part of a heterologous gene with respect to the transfected species. 13. Method according to one of the preceding claims, characterized in that the vector contains intervening sequences between the gene of the insertion and the flanking sequences. 14. Method according to claim 13, characterized in that the intervening sequences comprise a sequence that encodes a selection marker that allows the selection of the transformants and optionally a marker gene, such as the lacZ gene. 15. Method according to one of the preceding claims, characterized in that the transfection is carried out by means of electroporation. 16. Method according to one of the preceding claims, characterized in that the polymerase chain reaction (PCR) is used for the amplification of the DNA sequence of the locus at which the insertion is located, in order to check whether the insertion has taken place at the desired location. 17. A method according to any one of the preceding claims, characterized in that the DNA insert comprises part of an interferon gene, an interleukin gene, a receptor gene such as the retinoic acid, the β-3 adrenergic or the HIV receptor, or a disease-associated gene or the lacZ gene. 18. Method according to one of the preceding claims, characterized in that the DNA insert between the flanking sequences contains, on the one hand, a DNA sequence intended to be recombined with the complementary DNA in the recipient gene to generate a recombinant gene and, on the other hand, a sequence encoding a selectable marker allowing the selection of transformants and a promoter controlling the expression of the selectable marker. 19. A method for producing transgenic animals, characterized in that ES cells are transfected according to the method of any one of claims 1 to 18 and screened for the occurrence of homologous recombination to ensure correct integration of the foreign gene, the cells are injected into embryos which are at a stage where they are able to integrate the transfected cells, e.g. at the blastocyst stage, and which are subsequently reimplanted into a surrogate mother, and the chimeric individuals obtained at the end of pregnancy and in which colonization by ES cells derived from the germ are crossed to obtain transgenic animals that are heterozygous for the exchanged gene. 20. A nucleic acid sequence, namely the insertion sequence, containing: (i) a sequence comprising all or part of a gene which can be converted into a functional form or whose functionality can be increased when recombined with complementary DNA so as to form a complete recombinant gene in the genome of a eukaryotic cell, provided that if the part of the gene consists of a coding sequence, it is a coding sequence which does not originate from a gene encoding a selection marker, ii) two so-called "flanking" sequences on either side of the sequence mentioned under i), each of which is homologous to two genomic sequences located next to the desired insertion site in the recipient gene. 21. Nucleic acid sequence according to claim 20, characterized in that it contains, among other things, between the flanking sequences, a sequence coding for a selection marker which allows the selection of transformants and a promoter which controls the expression of the selection marker. 22. Nucleic acid sequence according to claim 20 or 21, characterized in that it contains, inter alia, a sequence which is eliminated by the process of homologous recombination and which allows the selection of transformants resulting from the event of homologous recombination. 23. Vector containing the nucleic acid sequence according to claim 20 or 21. 24. A nucleic acid sequence comprising the plasmid pGN as shown in Figure 1 and on either side of the pGN sequence two flanking sequences each homologous to two genomic sequences located adjacent to a desired insertion site in a recipient gene. 25. Eukaryotic cells transformed according to the method of claim 1, 2 or 7. 26. Cells according to claim 25, characterized in that they are ES cells. 27. Use of the nucleic acid sequence according to claim 20 for the modification of eukaryotic cells with a view to a functional or more efficient expression of the sequence or of an endogenous sequence. 28. Nucleic acid sequence comprising: (i) a regulatory sequence intended to be recombined with a complementary DNA to produce a complete recombinant gene in the genome of a eukaryotic cell, and ii) on either side of the sequence mentioned under i) two so-called "flanking" sequences, each of which is homologous to two genomic sequences located next to the desired insertion site in a recipient gene.